过滤 pandas 数据框，在 tkinter/pandastable 中显示

Question

我正在尝试创建一个涉及显示中型熊猫数据框的应用程序。我正在为我的用户界面使用 pandastable 和 tkinter。我希望用户能够做的部分事情是从组合框中选择一个过滤器，然后只显示匹配的行。

下面是我尝试使用的代码的简化但完整的版本。我创建了一个包含核心 UI 代码的类。最初一切看起来都很棒。我在我的类（change_df）中创建了一个测试按钮函数，它创建了一个虚拟列，效果很好。我在类（change_df_combo）中有另一个函数可以成功过滤数据框，但它不会显示在屏幕上。即，我可以打印 df 或导出它并且它已过滤。但是，这不会显示在屏幕上。

我假设我的班级或 tkinter 本身的循环不正确，但对于我的生活，我无法弄清楚！有什么想法吗？

from win32com.shell import shell, shellcon
import pandas as pd
import openpyxl
from openpyxl import load_workbook
import numpy as np
import time
from tkinter import *
from tkinter import ttk
from pandastable import Table, TableModel

class UserInterface(Frame):
    # Launch the df in a pandastable frame
    def __init__(self, parent=None):
        global ui_df
        global pt
        ui_df = pos_df
        self.parent = parent
        Frame.__init__(self)
        self.main = self.master
        #self.main.geometry('800x600+0+0')
        f = Frame(self.main)
        f.grid(column=0, row=1, sticky=(E, W))
        screen_width = f.winfo_screenwidth() * 0.8
        screen_height = f.winfo_screenheight() * 0.7
        self.table = pt = Table(f, dataframe=ui_df, height = screen_height, width = screen_width)
        pt.show()
        return

    def change_df(self, col_val_input):
        #Responds to button
        ui_df['Test col'] = col_val_input
        pt.show()

    def change_df_combo(self, event):
        #Responds to combobox, supposed to filter by 'Sec_type'
        combo_selection = str(combo_box.get())
        ui_df = pos_df[pos_df['Sec_type'] == combo_selection]
        pt.show()

#Create dataframe
pos_data = {'Location' : ['Denver', 'Boulder', 'Phoenix', 'Reno', 'Portland',
'Eugene', 'San Francisco'], 'Sec_type' : ['mbus', 'mbus', 'vmus', 'caus',
'vmus', 'mbus', 'mbus']}
pos_df = pd.DataFrame(data = pos_data)

#Launch Tkinter basics
root = Tk()
root.title("S test...")

mainframe = ttk.Frame(root, padding="3 3 12 12")
mainframe.grid(column=0, row=0, sticky=(N, W, E, S))
mainframe.columnconfigure(0, weight=1)
mainframe.rowconfigure(0, weight=1)
root.columnconfigure(0, weight=1)
root.rowconfigure(0, weight=1)

ui_display = UserInterface(mainframe)

#'Test' button, this works fine
col_val_input = 'It worked!'
test_button = ttk.Button(mainframe, text = 'Test', command= lambda: ui_display.change_df(col_val_input))
test_button.grid(column=0, row=0, sticky=(W))

#Combobox, works to just show choice, but not to filter
combo_choices = ['mbus', 'vmus', 'caus']
choice = StringVar()
combo_box = ttk.Combobox(mainframe, textvariable=choice)
combo_box['values'] = combo_choices
combo_box.grid(column=1, row=0, sticky=(W))
combo_box.bind('<<ComboboxSelected>>', ui_display.change_df_combo)

root.mainloop()

编辑添加： 将导入的 df 更改为玩具 df 以进行测试

Answer 1

好的，我想出了一些可行的方法。不确定它是否是有史以来写的最 Pythonic 的代码，但它是这样的：（只发布类段）

class UserInterface(Frame):
    # Launch the df in a pandastable frame
    def __init__(self, parent=None):
        global ui_df
        ui_df = pos_df
        self.parent = parent
        self.refresh_df(df = ui_df)

    def refresh_df(self, df):
        Frame.__init__(self)
        self.main = self.master
        f = Frame(self.main)
        f.grid(column=0, row=1, sticky=(E, W))
        screen_width = f.winfo_screenwidth() * 0.8
        screen_height = f.winfo_screenheight() * 0.7
        self.table = pt = Table(f, dataframe=df, height = screen_height, width = screen_width)
        pt.show()
        return

    def change_df(self, col_val_input):
        #Responds to button
        ui_df['Test col'] = col_val_input
        self.refresh_df(df=ui_df)

    def change_df_combo(self, event):
        #Responds to combobox, supposed to filter by 'Sec_type'
        combo_selection = str(combo_box.get())
        ui_df = pos_df[pos_df['Sec_type'] == combo_selection]
        ui_df['Test col combo'] = combo_selection
        self.refresh_df(df=ui_df)